When people interact with others, they control how much personal information they reveal, and, hence, make decisions about their privacy. Similarly, in online settings, they decide how much personal information they expose. However, online, their decisions might have greater consequences because of the persistent nature of disclosed information and technological complexities. Hence, technologies should provide people with comprehensive knowledge, enabling informed choices, and reducing privacy risks. Nevertheless, little is known about the interplay of the different factors that influence people's privacy choices. Sometimes, people over-disclose their personal information despite their privacy concerns, presumably relying on fast thinking of homo heuristicus. Yet, it is common to expect that people's decisions should result in more "rational" outcomes; this may decrease privacy risks - here, bringing to the forefront thinking of homo economicus.  This thesis presents insights into the understanding of online privacy decisions. Through a series of studies, we investigate the many factors that influence privacy-related attitudes and behaviors. Our work examines the interaction of the intrinsic (e.g., personality) and extrinsic (e.g., visual design of a privacy interface) determinants of online choices (e.g., disclosure). Overall, we demonstrate the complexity of human decisions in the context of online privacy, suggesting that people's choices are context-dependent and the borders between decisions of the homo heuristicus and economicus are blurry. Thus, we conclude that it is homo varius that can make privacy-aware choices by switching between rational calculations and heuristical thinking, depending on the interplay of different intrinsic and extrinsic determinants. This work provides a knowledge base for future studies investigating privacy decisions. Further, it contributes insights for privacy practitioners (e.g., designers) that may help improve current privacy designs.